Bench-to-bedside review: oxygen debt and its metabolic correlates as quantifiers of the severity of hemorrhagic and post-traumatic shock

Abstract

Evidence is increasing that oxygen debt and its metabolic correlates are important quantifiers of the severity of hemorrhagic and post-traumatic shock and and may serve as useful guides in the treatment of these conditions. The aim of this review is to demonstrate the similarity between experimental oxygen debt in animals and human hemorrhage/post-traumatic conditions, and to examine metabolic oxygen debt correlates, namely base deficit and lactate, as indices of shock severity and adequacy of volume resuscitation. Relevant studies in the medical literature were identified using Medline and Cochrane Library searches. Findings in both experimental animals (dog/pig) and humans suggest that oxygen debt or its metabolic correlates may be more useful quantifiers of hemorrhagic shock than estimates of blood loss, volume replacement, blood pressure, or heart rate. This is evidenced by the oxygen debt/probability of death curves for the animals, and by the consistency of lethal dose (LD)25,50 points for base deficit across all three species. Quantifying human post-traumatic shock based on base deficit and adjusting for Glasgow Coma Scale score, prothrombin time, Injury Severity Score and age is demonstrated to be superior to anatomic injury severity alone or in combination with Trauma and Injury Severity Score. The data examined in this review indicate that estimates of oxygen debt and its metabolic correlates should be included in studies of experimental shock and in the management of patients suffering from hemorrhagic shock.

Figure 1

Probability of death as a function of oxygen debt. (a) Regression-derived relation of Kaplan–Meier probability of death as a function of increasing oxygen debt (O₂D) in a canine O₂D hemorrhagic shock model. Noted on the figure are the O₂D values for lethal dose (LD)₂₅ (i.e. a dose sufficient to kill 25% of the population studied), LD₅₀, and LD₇₅ probabilities. Points plotted along the regression line and its 95% confidence limits represent the actual Kaplan–Meier survival (S) values at 60 min of hemorrhage, or values at the time of death (D) for nonsurviving animals dying during the hemorrhage period or within 5 min of the 60 min hemorrhage sample. Note the good correlation of Kaplan–Meier points to the regression-estimated line. Reproduced with permission from Dunham and coworkers [9]. (b) Probability of death as a function of O₂D in a pig O₂D hemorrhagic shock model. Noted on the figure are the O₂D values for LD₂₅, LD₅₀, and LD₇₅ probabilities. Points plotted along the regression line and its 95% confidence limits represent the values of cumulative O₂D (in ml/kg) at 60 min of hemorrhage for survivors (marked with circles) and nonsurvivors (marked with squares). Modified from Rixen and coworkers [44].

Figure 2

Mortality as a function of base deficit. Mortality curves presented as a function of the admission base deficit in more than 8000 multiple trauma patients derived from four independent studies. Modified from Zander [89].

Figure 3

Interaction between base excess, Glasgow Coma Scale (GCS) and mortality. (a) Linear logistic model for predicting mortality from GCS and admission extracellular base excess (BEA) for 185 patients with blunt traumatic hepatic injury (λ = - 0.21 [GCS] - 0.147 [BEA_ECF] + 0.285; P < 0.0001 for model). (b) Predicted versus observed mortality in linear logistic model from GCS and BEA for patients with blunt traumatic hepatic injury. ECF, extracellular fluid. Reproduced with permission from Siegel and coworkers [29].

Figure 4

Discrimination and calibration of the multivariate outcome prediction model. (a) Discrimination. Receiver operating characteristic curve of the multivariate outcome prediction model based on base deficit, Glasgow Coma Scale score, prothrombin time, age and Injury Severity Score, compared with that derived from the Trauma and Injury Severity Score (TRISS) in the validation set of 1745 multiple trauma patients. The diagonal line corresponds to a test that is sensitive or specific just by chance. The area under the curve for the multivariate outcome prediction model is 0.901 and that for the TRISS score is 0.866. (b) Calibration. Predicted versus observed mortality for the multivariate outcome prediction model and the TRISS score in the validation set of 1745 multiple trauma patients.